Proa Analytics is proud to lead a 2 year ARENA funded project demonstrating its advanced, solar forecasting system at three Australian solar farms.
These fluctuations impact the power system, so improved forecasts will help to integrate solar generation into the grid and reduce the cost of solar generation. This forecasting project will forecast the generation of three large scale solar farms using a new solar forecasting system developed by Australian company Proa Analytics. Every five minutes, the Proa Forecasting System (PFS) will produce an updated generation forecast for the next 5-minute period for each solar farm, as well as longer forecasts up to several days ahead. Proa’s forecasts use an innovative combination of four different technologies: skycams, satellites, live data and numerical simulations.
The Proa Forecasting System (PFS) was developed in Australia for Australian conditions. It consists of four individual solar forecasting techniques that are combined to produce an optimal forecast. These include three advanced and proprietary forecasting techniques developed by Proa Analytics — our methods of geostationary satellite cloud motion vectoring (CMV) algorithms, skycam CMV, and live data techniques. While each technology performs best at different time scales and weather conditions, this project aims to demonstrate the benefits of optimally combining them to maximise forecast accuracy.
The project will also demonstrate new methods of identifying clouds at night using infrared satellite images to make accurate solar forecasts in the hours before dawn, and use an infrared skycam to provide additional information on cloud structure during the day.
Improved solar forecasts are required to help AEMO operate the NEM with greater amounts of solar generation and help additional solar generation to enter the grid.
The project will also analyse the costs and benefits of different forecasting technologies and combinations of technologies and publish this information to the market, to help solar farms to determine the best choice of forecasting technologies that suit their technology type, size, and location.